                 Changes in version 2.0.1 (2026-09-07)                  

  - Removed all contributed runtime dependencies. Only standard R
    libraries remain in Imports; testthat, knitr and rmarkdown are
    development suggestions.
  - Added internal mapping, dummy encoding, moments, shorth mode,
    entropy, distance and forecast metric calculations.
  - Replaced DTW with an internal symmetric2 dynamic-programming
    recurrence.
  - Replaced future/furrr with standard parallel PSOCK workers and
    guaranteed worker cleanup on exit.
  - Replaced digest hashing with a bounded, single-entry exact-input
    cache.
  - Plots now use base graphics through plot()/print() methods for
    segen_plot objects. Adding ggplot2 layers is no longer supported.
  - Numeric imputation now uses linear interpolation with constant
    endpoints; categorical gaps use the most frequent level. Loess uses
    span 0.75 and degree 1. These are intentional behavioral changes
    from 2.0.0.
  - Binary entropy now uses empirical category frequencies. Percentage
    errors use percentage units; near-zero metric denominators are
    bounded at 1e-8.
  - Integrated the point forecast before combining it with integrated
    draws.
  - Updated vignette, namespace, dataset documentation and regression
    tests.

                 Changes in version 2.0.0 (2025-08-19)                  

  - Added the segen-introduction vignette with numeric, reproducibility,
    and optional-method examples.
  - Added early validation for integer controls, error scaling options,
    and parallel worker counts.
  - Replaced tictoc, readr, and scales usage with base R equivalents.
  - Moved DTW and parallel execution packages to Suggests; sequential
    execution remains the default.
  - Strengthened distance-cache keys so distinct input matrices cannot
    collide merely because they share dimensions, mean, and standard
    deviation.
  - Added tests for validation and cache correctness.
  - Removed the mandatory philentropy dependency and implemented the
    categorical distance calculations internally.
  - Removed stale namespace imports for optional packages and utility
    packages; DTW and parallel execution are accessed only when
    explicitly requested.
  - Repaired package metadata so devtools::document() does not
    reintroduce removed dependencies.

                        Changes in version 1.2.0                        

  - Fixed mixed-type detection bug (class_index now per-column; strict
    all-numeric OR all-categorical check).

  - Corrected best_deriv() p-values length/off-by-one and added
    short-series guard.

  - Patched doxa_filter() binary clamp (now clamps to [0,1] instead of
    < 1 typo).

  - Hardened metrics: added safe_scale() to avoid zero/NA scale; fixed
    benchmark to use mean(actuals) for "average".

  - Validated inputs (ci, windows, samples, dates length) and added
    early error for too-short series.

  - Stabilized weight rescaling/min-max with zero-range handling;
    vectorized ensemble normalization.

  - Improved date handling for plots/outputs (consistent step; correct
    future index).

  - Reduced brittleness/perf costs (preallocation where useful, safer
    splitting, fewer coercions) and kept seed propagation for
    reproducibility.

  - Added compute_distance() with memoized caching and new methods:
    "correlation" (1−corr) and "dtw" (guarded on dtw pkg).

  - Replaced Gaussian weight sampling with kernel weights vs. the most
    recent window + Dirichlet draws for uncertainty (faster, stabler).

  - Implemented conformal prediction: CI bounds now calibrated from
    rolling residuals; plot uses these bounds at the requested ci.

  - Introduced optional parallelism (use_parallel, parallel_workers)
    using furrr/future with seeded execution; falls back to sequential
    if unavailable.

  - Enabled distance-matrix caching (via digest when present) for reuse
    across samples and to cut repeated computation.

  - Kept API backward-compatible; new options default off and legacy
    distance methods unchanged.

                 Changes in version 1.1.0 (2022-08-15)                  

  - Rationalization and streamlining of processes.

                 Changes in version 1.0.0 (2022-06-28)                  

  - Added a NEWS.md file to track changes to the package.
